Novel inflammatory biomarkers enhance prediction of sepsis-associated acute kidney injury: a machine learning approach

Abstract

Background: Sepsis-associated acute kidney injury (S-AKI) is a common complication of sepsis, and early identification can improve patient prognosis. This study incorporated novel inflammatory markers as features to construct a model and employed 6 machine learning methods to predict the occurrence of S-AKI.

Methods: A total of 3613 patients with sepsis were included in this study. Novel inflammatory markers, including neutrophil-to-lymphocyte ratio, monocyte-to-lymphocyte ratio, systemic inflammatory response index, systemic immune-inflammation index, systemic inflammatory aggregate index, lactate-to-albumin ratio, and prognostic nutritional index, along with demographic characteristics, clinical conditions, and routine laboratory results, were used to construct the model. The machine learning methods employed included logistic regression, support vector machine, random forest (RF), extreme gradient boosting (XGBoost), and ensemble methods(RF+XGBoost). Model performance and stability were evaluated using 5-fold cross-validation. Model performance was assessed using the area under the receiver-operating characteristic curve, sensitivity, specificity, accuracy, precision, recall, and F1 score. Additionally, SHapley Additive exPlanations values were used to interpret the predictive model.

Results: In the final algorithm group, the ensemble model of RF and XGBoost (0.843; 95% confidence interval: 0.820-0.866) was higher than those of other models. Among the single models, the XGBoost model exhibited the highest sensitivity (0.856) and F1 score (0.780), indicating its stronger ability to identify patients who will develop S-AKI, albeit at the expense of lower specificity (0.667). The 4 most influential features for XGBoost were mechanical ventilation, mean arterial pressure, blood urea nitrogen level, and sequential organ failure assessment score. Among the 3 novel inflammatory markers, lactate-to-albumin ratio showed the greatest effect.

Conclusion: We successfully developed machine learning methods to predict S-AKI, highlighting the importance of novel inflammatory markers in model construction. This breakthrough offers novel perspectives for feature selection in the future development of related predictive models.

Keywords

Acute kidney injury / Artificial intelligence / Machine learning / Prediction / Sepsis

Conflict of interest statement

The authors declare no conflict of interest.

Author contributions

Zhang Y was responsible for acquiring the database, data extraction, curation, validation, and wrote the original draft of the manuscript. Han X conceived, designed, and supervised the study. Lu S and Zeng Y performed formal analysis and investigation. Zhou Z conducted statistical analyses and created visualizations. Xu X and Fan M assisted in methodology development and project administration. Yan X and Zhu Y reviewed and edited the manuscript, and provided critical intellectual content. All authors have reviewed and approved the final manuscript.

Funding

This study was funded by National Natural Science Foundation of China (82550116), Hunan Provincial Natural Science Foundation of China (2024JJ2038 and 2024JJ9161), Central Government Guides Local Science and Technology Development Fund Projects (2024ZYC031), Hunan Health High-Level Talent Project (R2023073), National Key Clinical Specialty Scientific Research Project (Z2023114), Research Project of the Health Commission of Hunan Province (20256566), Young Doctor Foundation of Hunan Provincial People’s Hospital (BSJJ202209), and Key Cultivation Project of Hunan Provincial People’s Hospital (RS2022A06).

Ethical approval of studies and informed consent

All the data were sourced from the MIMIC-IV database. Zhang Y extracted data after completing online training (certification number 64967484) and passing the exam on the website. The ethics activities of MIMIC-IV have been approved.

Acknowledgements

This study utilized the MIMIC-IV database (Medical Information Mart for Intensive Care), which is provided by the Computer Science and Artificial Intelligence Laboratory (CSAIL) at the Massachusetts Institute of Technology (MIT) and funded by the National Institutes of Health (NIH). We would like to express our sincere gratitude to the MIT CSAIL team, as well as to all the researchers and support staff involved in this project, for providing such invaluable data resources for medical research. Additionally, we extend our thanks to all the contributors and maintainers of the MIMIC database, whose hard work has ensured the ongoing updates and availability of this resource.

References

[1] Singer M, Deutschman CS, Seymour CW, et al. The third international consensus definitions for sepsis and septic shock (Sepsis-3). JAMA. 2016; 315(8):801-810. doi:10.1001/jama.2016.0287

[2] Zarbock A, Nadim MK, Pickkers P, et al. Sepsis-associated acute kidney injury: consensus report of the 28th acute disease quality initiative workgroup. Nat Rev Nephrol. 2023; 19(6):401-417. doi:10.1038/s41581-023-00683-3

[3] Kellum JA, Chawla LS, Keener C, et al. ProCESS and ProGReSS-AKI Investigators. The effects of alternative resuscitation strategies on acute kidney injury in patients with septic shock. Am J Respir Crit Care Med. 2016; 193(3):281-287. doi:10.1164/rccm.201505-0995OC

[4] de Geus HR, Betjes MG, Bakker J. Biomarkers for the prediction of acute kidney injury: a narrative review on current status and future challenges. Clin Kidney J. 2012; 5(2):102-108. doi:10.1093/ckj/sfs008

[5] Tseng PY, Chen YT, Wang CH, et al. Prediction of the development of acute kidney injury following cardiac surgery by machine learning. Crit Care. 2020; 24(1):478. doi:10.1186/s13054-020-03179-9

[6] Zhang L, Wang Z, Zhou Z, et al. Developing an ensemble machine learning model for early prediction of sepsis-associated acute kidney injury. iScience. 2022; 25(9):104932. doi:10.1016/j.isci.2022.104932

[7] Zhang Q, Song J, Wu H, et al. Intratumoral microbiota associates with systemic immune inflammation state in nasopharyngeal carcinoma. Int Immunopharmacol. 2024; 141:112984. doi:10.1016/j.intimp.2024.112984

[8] Rödel APP, Fernandes YM, Brisolara JV, De Carvalho JAM, Moresco RN. Role of preoperative inflammatory blood cell indexes as a postoperative risk predictor among patients undergoing on-pump cardiac surgery. Int J Lab Hematol. 2025; 47(1):87-92. doi:10.1111/ijlh.14390

[9] Tahavvori A, Mosaddeghi-Heris R, Ghanbari Sevari F, et al. Combined systemic inflammatory indexes as reflectors of outcome in patients with COVID-19 infection admitted to ICU. Inflammopharmacology. 2023; 31(5):2337-2348. doi:10.1007/s10787-023-01308-8

[10] Djordjevic D, Rondovic G, Surbatovic M, et al. Neutrophil-to-lymphocyte ratio, monocyte-to-lymphocyte ratio, platelet-to-lymphocyte ratio, and mean platelet volume-to-platelet count ratio as biomarkers in critically ill and injured patients: which ratio to choose to predict outcome and nature of bacteremia? Mediators Inflamm. 2018; 2018:3758068. doi:10.1155/2018/3758068

[11] Shin J, Hwang SY, Jo IJ, et al.; Korean Shock Society (KoSS) Investigators. Prognostic value of the lactate/albumin ratio for predicting 28-day mortality in critically ill sepsis patients. Shock. 2018; 50(5):545-550. doi:10.1097/SHK.0000000000001128

[12] Chen Q, Zhan H, Chen J, Mo J, Huang S. Predictive value of lactate/albumin ratio for death and multiple organ dysfunction syndrome in patients with sepsis. J Med Biochem. 2024; 43(4):617-625. doi:10.5937/jomb0-46947

[13] Sun B, Chen Y, Man Y, Fu Y, Lin J, Chen Z. Clinical value of neutrophil-to-lymphocyte ratio and prognostic nutritional index on prediction of occurrence and development of diabetic foot-induced sepsis. Front Public Health. 2023; 11:1181880. doi:10.3389/fpubh.2023.1181880

[14] Kyo D, Tokuoka S, Katano S, et al. Comparison of nutrition indices for prognostic utility in patients with sepsis: a real-world observational study. Diagnostics (Basel). 2023; 13(7):1302. doi:10.3390/diagnostics13071302

[15] Johnson AEW, Bulgarelli L, Shen L, et al. MIMIC-IV, a freely accessible electronic health record dataset. Sci Data. 2023; 10(1):1. doi:10.1038/s41597-022-01899-x

[16] LaValley MP. Logistic regression. Circulation. 2008; 117(18):2395-2399. doi:10.1161/CIRCULATIONAHA.106.682658

[17] Hu J, Szymczak S. A review on longitudinal data analysis with random forest. Brief Bioinform. 2023; 24(2):bbad002. doi:10.1093/bib/bbad002

[18] Wang H, Shao Y, Zhou S, Zhang C, Xiu N. Support vector machine classifier via L(0/1) soft-margin loss. IEEE Trans Pattern Anal Mach Intell. 2022; 44(10):7253-7265. doi:10.1109/TPAMI.2021.3092177

[19] Sutou A, Wang JF. Influence-balanced XGBoost: improving XGBoost for imbalanced data using influence functions. IEEE Access. 2024; 12:193473-193486. doi:10.1109/access.2024.3520159

[20] Watchorn J, Huang D, Bramham K, Hutchings S. Decreased renal cortical perfusion, independent of changes in renal blood flow and sublingual microcirculatory impairment, is associated with the severity of acute kidney injury in patients with septic shock. Crit Care. 2022; 26(1):261. doi:10.1186/s13054-022-04134-6

[21] Liu AB, Tan B, Yang P, et al. The role of inflammatory response and metabolic reprogramming in sepsis-associated acute kidney injury: mechanistic insights and therapeutic potential. Front Immunol. 2024; 15:1487576. doi:10.3389/fimmu.2024.1487576

[22] Gupta A, Berg DT, Gerlitz B, et al. Role of protein C in renal dysfunction after polymicrobial sepsis. J Am Soc Nephrol. 2007; 18(3):860-867. doi:10.1681/ASN.2006101167

[23] Han SJ, Lee HT. Mechanisms and therapeutic targets of ischemic acute kidney injury. Kidney Res Clin Pract. 2019; 38(4):427-440. doi:10.23876/j.krcp.19.062

[24] Liu S, Hartleben B, Kretz O, et al. Autophagy plays a critical role in kidney tubule maintenance, aging and ischemia-reperfusion injury. Autophagy 2012; 8(5):826-837. doi:10.4161/auto.19419

[25] Bellomo R, Kellum JA, Ronco C, et al. Acute kidney injury in sepsis. Intensive Care Med. 2017; 43(6):816-828. doi:10.1007/s00134-017-4755-7

[26] Nusshag C, Wei C, Hahm E, et al. suPAR links a dysregulated immune response to tissue inflammation and sepsis-induced acute kidney injury. JCI Insight. 2023; 8(7):e165740. doi:10.1172/jci.insight.165740

[27] Xiao W, Lu Z, Liu Y, et al. Influence of the initial neutrophils to lymphocytes and platelets ratio on the incidence and severity of sepsis-associated acute kidney injury: a double robust estimation based on a large public database. Front Immunol. 2022; 13:925494. doi:10.3389/fimmu.2022.925494

[28] Gocze I, Koch M, Renner P, et al. Urinary biomarkers TIMP-2 and IGFBP7 early predict acute kidney injury after major surgery. PLoS One 2015; 10(3):e0120863. doi:10.1371/journal.pone.0120863

[29] Bihorac A, Chawla LS, Shaw AD, et al. Validation of cell-cycle arrest biomarkers for acute kidney injury using clinical adjudication. Am J Respir Crit Care Med. 2014; 189(8):932-939. doi:10.1164/rccm.201401-0077OC

[30] Antonelli M, Azoulay E, Bonten M, et al. Year in review in intensive care medicine 2010: I. acute renal failure, outcome, risk assessment and ICU performance, sepsis, neuro intensive care and experimentals. Intensive Care Med. 2011; 37(1):19-34. doi:10.1007/s00134-010-2112-1

[31] Tutunea-Fatan E, Arumugarajah S, Suri RS, et al. Sensing dying cells in health and disease: the importance of kidney injury molecule-1. J Am Soc Nephrol. 2024; 35(6):795-808. doi:10.1681/ASN.0000000000000334

[32] Xu L, Wang X, Pu P, Li S, Shao Y, Li Y. Ultrasonic image features under the intelligent algorithm in the diagnosis of severe sepsis complicated with renal injury. Comput Math Methods Med. 2022; 2022:2310014. doi:10.1155/2022/2310014

[33] Wu Z, Liu WQ, Tang L, et al. Lactate-mitochondrial crosstalk: a new direction in the treatment of sepsis-induced acute kidney injury. Cell Biol Int. 2024; 48(11):1621-1624. doi:10.1002/cbin.12240

[34] Durand F, Francoz C, Asrani SK, et al. Acute kidney injury after liver transplantation. Transplantation. 2018; 102(10):1636-1649. doi:10.1097/TP.0000000000002305

[35] Tu H, Su J, Gong K, et al. A dynamic model to predict early occurrence of acute kidney injury in ICU hospitalized cirrhotic patients: a MIMIC database analysis. BMC Gastroenterol. 2024; 24(1):290. doi:10.1186/s12876-024-03369-7

[36] Fang Y, Zhang Y, Zhang X. The elevated lactate dehydrogenase to albumin ratio is a risk factor for developing sepsis-associated acute kidney injury: a single-center retrospective study. BMC Nephrol. 2024; 25(1):201. doi:10.1186/s12882-024-03636-5

[37] Shi X, Zhong L, Lu J, Hu B, Shen Q, Gao P. Clinical significance of the lactate-to-albumin ratio on prognosis in critically ill patients with acute kidney injury. Ren Fail. 2024; 46(1):2350238. doi:10.1080/0886022X.2024.2350238

[38] Poston JT, Koyner JL. Sepsis associated acute kidney injury. BMJ. 2019;364:k4891. doi:10.1136/bmj.k4891